The present study aimed to map and analyze the spatiotemporal dynamics of flood hazard in the Kalarud watershed, Babol (northern Iran), covering an area of 216 km². This basin, with an elevation range of 45–2272 m and an average slope of 19.46%, is among the flood-prone zones of the Central Alborz Mountains. To achieve a comprehensive assessment, a hybrid approach combining Shannon Entropy (SE), Evidential Belief Function (EBF), and Markov Chain (MC) models was applied to evaluate both spatial and temporal aspects of flood risk.

Nine conditioning factors—elevation, slope, slope aspect, drainage density, distance from rivers and roads, land use, soil, and Topographic Wetness Index (TWI)—were analyzed in a GIS environment. Results of the SE model indicated that elevation (0.1983), slope aspect (0.1517), and slope (0.1423) were the most influential factors. The EBF model confirmed the 48% cumulative weight of these variables, while the Markov Chain model was employed to forecast spatial changes in flood hazard from 2013 to 2030.

Model performance evaluation showed that EBF (AUC = 0.83, RMSE = 0.219) outperformed the Shannon model (AUC = 0.71, RMSE = 0.293), whereas the Markov model (AUC = 0.79) exhibited stable temporal behavior. Overall, 26.4% of the basin area falls within high and very high hazard classes, mainly encompassing the central and southern sectors, including the villages of Shiyadeh, Anjilek, and Lamsukola. The integrated EBF–Markov–SE framework effectively addresses spatial uncertainty and temporal dynamics, providing a reliable tool for flood risk management, watershed planning, and sustainable development in northern Iran.